Abstract:

Rough set theory is a new soft calculation method which used to process fuzzy and indetermination problems. Edge is a kind of collection that involves big grayscale changes points. In nonnoise jamming’s situation, this is a characteristic that distinguishies the difference between edge and nonedge. But in the practical application, the image unavoidably mixes with some noise which impact accuracy of edge detection. This paper presents an image edge detection theory based on rough sets. According to the set’s similar relations of rough set, it first finds possible edge point set, using the characteristic of big grayscale changes. Then, uses the difference between noise and edge to find noise set. Finally, the difference between the two sets is on the verge of final demand points. Experimental results show,compared to traditional methods of edge detection, this method has certain improvement in examination accuracy.

Key words: rough sets, edge detection, edge gradient